INDEX

Explanations

phrases identifying causes of negative events or consequences

oai_token-act-pair · gpt-3.5-turbo

phrases indicating effects or results of events

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Technique

-0.77

 scrimmage

-0.70

 motto

-0.68

 skelet

-0.63

ault

-0.63

aeper

-0.62

halla

-0.62

 Leopard

-0.61

 Fargo

-0.60

 request

-0.59

POSITIVE LOGITS

 cele

1.03

netflix

0.93

 havoc

0.92

uria

0.85

 irre

0.82

 irreversible

0.78

parable

0.74

auga

0.73

 blindness

0.70

 cancell

0.69

Activations Density 0.020%

phrases identifying causes of negative events or consequences

phrases indicating effects or results of events

No Comments

No Known Activations